Analysis of time-course gene expression profiles to study regulation of cell growth in fed-batch bioreactors

Compared to other systems biology tools, genomic microarrays represent a mature platform that allows for facile access to the internal biological mechanisms of cell culture. While the large datasets generated by microarrays are a potential goldmine of information, ironically, it is their large size that also makes data-mining a cumbersome task. This can get further complicated by unavoidable batch effects generated when different datasets are combined. Furthermore, gene expression profiles are dependent on combinations of various complex intracellular events and as such identifying the signals primarily related to the phenotype of interest poses a substantial challenge. In this study we addressed these issues by utilizing a workflow that allows adjustment of time-course gene expression datasets for batch effects and incorporates the use of sparse partial least squares analysis to identify specific genes of interest. We were able to identify a set of relevant genes that displayed a strong correlation with cell growth in fed-batch bioreactors under different nutrient compositions. By conducting further biological network analysis, we identified four transcriptional regulators, namely ATP7B, SREBP1, SCAP and INSIG2 that are responsible for regulation of these genes and are likely important drivers for cell growth differences in response to change in nutrient composition.